AI agents raise new data governance challenges for businesses
The incorporation of artificial intelligence agents into business processes poses new challenges for data management, metric interpretation and the oversight of automated actions.
Data management, understanding business processes and overseeing automated actions are among the factors organizations must consider when incorporating artificial intelligence into their operations. Against this backdrop, business analytics platforms are beginning to integrate capabilities that enable AI agents to work with contextualized information and perform tasks within corporate workflows.
One such tool is Zoho Analytics 7.0, the latest version of Zoho Corporation’s data analytics and business intelligence platform. It introduces a multi-agent architecture, models for interpreting business context, and governance and oversight mechanisms.
Announced on October 28, 2026, the update includes features for querying information, generating analyses and triggering actions based on an organization’s data.
The platform incorporates two layers to structure the information used by AI systems. The semantic layer functions as a glossary that standardizes business descriptions, instructions and metrics, with the aim of ensuring consistency across queries and analyses.
The ontological layer, in turn, represents business units and their relationships, enabling agents to interpret data within each organization’s operational structure.
With this model, agents’ responses and recommendations can be built on definitions and relationships specific to each company, allowing queries and analyses to be contextualized according to its operational characteristics.
”We are at the beginning of the great leap in analytics tools, where a solid data foundation is the starting point for all agentic activity in a company’s operations. Agents do not work with data alone: they need a robust intelligence layer. The new features in Zoho Analytics provide an open, extensible and reliable layer so that every AI tool has the context, metrics and governance needed to deliver reliable answers,” said Raju Vegesna, Chief Evangelist at Zoho.
Agents for analysis and task execution.
One of the key components of the update is Ask Zia, the platform’s conversational assistant, which now operates on a multi-agent architecture comprising six specialized sub-agents. Their capabilities include data analysis and modeling, visualization, data preparation, programming, action execution and SQL query generation.
The tool retains conversation history and agent memory. It also allows organizations to choose between Zoho’s language model and integrating other artificial intelligence models, according to their preferences.
Another option is to develop custom agents using proprietary business information, semantic models and knowledge bases. These agents can connect to Model Context Protocol (MCP) servers, web search capabilities and other tools, both within and outside the system.
The platform also allows agents to execute actions and trigger workflows through an automation builder and contextual triggers. This enables analytical results to be linked to specific operational tasks rather than limiting their use to information retrieval and report generation.
MCP integration extends these capabilities beyond the Zoho environment. Developers and external agents can access Analytics capabilities through remote and local MCP servers, a command-line interface (CLI) and a plugin for working with Claude.
White-label agents can also be embedded in applications, portals and products from other organizations, with their respective permissions and governance controls.
AI oversight and governance.
Agent management is another key component of the update. The new version introduces a centralized catalog that brings together data, metrics, analyses and models, along with tools to track their origins, assess the impact of changes, and manage policies governing access to and use of information.
It also includes AI observability features that allow organizations to monitor which agents have been executed, who has used them, what information they have accessed and what responses they have generated. Administrators can review AI usage and costs by team, as well as configure Ask Zia’s behavior through administrative settings.
These capabilities are complemented by a Sandbox environment for developing, testing and deploying analytics solutions and agents before they go into production. Teams can revert to previous versions if they identify problems, facilitating controlled change management.
In terms of connectivity, the update adds more than 30 new connectors to the over 500 previously available. Integrations include Zoho ERP, Zoho Spend, Zoho FSM, Zoho Shifts, Zoho Voice, SAP, PayPal, Adobe Commerce, Twilio, Notion, Intercom, Supabase, Amazon DocumentDB and AlloyDB.
Applications across different business areas.
The new features are designed to serve different roles within organizations. Executives and business managers can query metrics through Ask Zia; finance and operations teams can work with definitions specific to their respective areas; and data analysts and scientists have additional tools to develop models and examine information.
For data analysts and scientists, the new version introduces more than 10 visualizations, scenario analysis, correlation matrices and the VARX forecasting model, as well as a data science environment with Jupyter notebooks. It also includes custom visual reports, pixel-perfect reports, PMML import and a redesigned version of Ask Zia for mobile devices.
Zoho Analytics is used by more than 100,000 companies worldwide and can be deployed in Zoho’s cloud, on-premises, through public cloud services such as AWS, Azure and Google Cloud, or in hybrid environments.
The new features are already available in the cloud version. Their rollout to hybrid and on-premises editions is scheduled for the first quarter of 2027.
With this update, Zoho expands its analytics platform’s capabilities to include the creation, integration and oversight of artificial intelligence agents. The platform brings together tools to connect business data, interpret information according to each organization’s context, and link analytical results to actions and workflows.
|
|
